LWN: Comments on "Ubuntu Touch, three years later"

Ubuntu Touch, three years later

knitzsche — Tue, 29 Dec 2015 16:11:59 +0000

Also note that the weather data in the Today scope is for your particular location (unless you choose to disable location data).

BQ’s Ubuntu smartphones now available globally

rigo — Thu, 24 Dec 2015 14:30:50 +0000

"The Nexus 4 is certainly past its prime at the end of 2015, but it still functions as a credible Android device. It is, in any case, the only phone handset on the list of supported devices other than the three that were sold (in locations far from your editor's home) with Ubuntu Touch pre-installed"

BQ’s Ubuntu smartphones now available globally

BQ store delivery information

This is just silly.

hummassa — Mon, 21 Dec 2015 03:34:45 +0000

It's not failure if it does something you wanted it to do.

For instance, I can want/need a safer/privacy-enabled phone.

You don't need to "be competitive", you just need to do what you want to do. I have homebrew raspberry-pi based home theatre boxes in all of my home. They are (each!) one of a kind, and they are astounding successes to me -- I can watch movies, netflix, and other stuff on them and I learnt how to assemble them and what software put on them and how to configure one. All objectives that I had for them are fulfilled.

Ubuntu Touch, three years later

jcm — Mon, 21 Dec 2015 02:42:27 +0000

The biggest question remains "why?"

As in "why does this distribution exist?". What is the purpose of Ubuntu on a cellphone? Why should someone seek to pre-install it? Competition is a great thing. But only if you can be competitive. Those who are trying to do phone Operating Systems who aren't Apple, Google, maybe Microsoft, and perhaps some well funded quasi-Chinese-govt-connected are all doomed to failure. Now, being utterly doomed to commercial failure can be a reasonable thing if you're trying to advance some other philanthropic cause. But what if you're just trying to be another also ran alternative to the OSes people actually want? In that case, it's time to give up now.

"Developing economies" are often the kinds of words used in response. But here's the rub. Folks in "developing economies" don't want a cheaper alternative to Android. They want exactly Android. Or iOS. They want what the rest of the world has, not what we think is good for them. So, I would say, if the purpose of Ubuntu on a phone is to help advance Free Software or be great for a small set of developers who will use it, then great. But if its continued existence is with the goal of commercial success then there's a memo or two that hasn't been read a long, long time ago in a galaxy far, far away about trying to go up against Apple, Google, and Microsoft, with less than billions of dollars in investment, incentives, and other ammunition.

Ubuntu Touch, three years later

jospoortvliet — Sun, 20 Dec 2015 15:58:56 +0000

Sure, I won't say cash alone is enough. You need drive, will, manage and motivate people and so on. But without cash, all that is rather hard... Not impossible, but hard.

Ubuntu Touch, three years later

khim — Sun, 20 Dec 2015 13:48:26 +0000

I'm not talking about cash. Apple has billions in cash, Microsoft has billions in cash, Google has billions in cash. Big deal. Cash alone wouldn't give you a brilliant design years ahead of competitors.

To do THAT you need to basically pull “brightest and smartest” from all around your company, convince them to work 14 hours a day somehow and constantly track and remove all the obstacles from their path. That's how iPhone arrived about two years ahead of Android, e.g.

Tim Cook is brilliant manager, but he couldn't do something like that. He couldn't even understand why would he NEED to do something like that. Which means that this time Apple will create a working device more-or-less simultaneously with Google—which would mean that they'll lose.

Ubuntu Touch, three years later

jrjohansen — Sat, 19 Dec 2015 23:15:19 +0000

>> slow response - partially, there are fixes in the stable channel that are not in the unstable channel.
>
> And that doesn't strike you as a particularly bad idea? There's a reason why, for example, the Linux kernel
> stable rules are what they are.

Not my choice, just stating a fact. I think its naming is unfortunate because that channel also has different semantics than debian unstable and yet people keep making that mistake.

> With regard to features, I think it's the right thing to concentrate on those that are available to LWN readers
> from the official development channels, and come with source code. When device vendors decide to ship their
> proprietary stuff on top of that, I do not consider that to be a part of Ubuntu Touch. I cannot even count on the
> continued availability of it. When inevitably the vendor loses interest in the device, I'm left with only those
> components that came with source code. If that's not enough to actually use the device, the device is now
> useless. I learned that the hard way. And yes, this also applies to most, if not all, Android devices.

sure, but again that was only part of the features. Unstable is under going the transition to gcc 5 and there have been some issues with the transition. I don't know all the reasons, I just know that unstable at the moment is a really not the channel to be using except for specific development testing.

Ubuntu Touch, three years later

lsl — Sat, 19 Dec 2015 20:36:30 +0000

> slow response - partially, there are fixes in the stable channel that are not in the unstable channel.

And that doesn't strike you as a particularly bad idea? There's a reason why, for example, the Linux kernel stable rules are what they are.

With regard to features, I think it's the right thing to concentrate on those that are available to LWN readers from the official development channels, and come with source code. When device vendors decide to ship their proprietary stuff on top of that, I do not consider that to be a part of Ubuntu Touch. I cannot even count on the continued availability of it. When inevitably the vendor loses interest in the device, I'm left with only those components that came with source code. If that's not enough to actually use the device, the device is now useless. I learned that the hard way. And yes, this also applies to most, if not all, Android devices.

Ubuntu Touch, three years later

jospoortvliet — Sat, 19 Dec 2015 17:43:19 +0000

Not sure if jobs would've strained Apple's resources. They have 100 billion I.cash - good luck spending that.

Ubuntu Touch, three years later

rsidd — Sat, 19 Dec 2015 15:29:53 +0000

Isn't it pretty common -- not just in phone software, but in free software generally -- for the dev channel to have debugging features enabled that are disabled in release channels? That alone would explain performance issues. I don't know if that's the case here but have run into it often enough.

Ubuntu Touch, three years later

warmcat — Sat, 19 Dec 2015 03:15:56 +0000

No... "here's what happened when I put it on device X" is exactly the kind of article needed and demonstrates the right amount of research right there.

Ubuntu Touch, three years later

ibukanov — Fri, 18 Dec 2015 21:55:32 +0000

> Too late. The war is already lost.

Well, there is an always a niche. 1% of billion users is still a big number. However niches require that vendors identifies the target group and work to support its interests. This is just not the case for Ubuntu touch or Tizen. They want to be a generic smarphone and so they indeed lost before even starting. Mozilla with Firefox OS at least tried to focus on a particular niche like low end smartphones, but it was executed badly and now days Android runs rather OK on phones under 70USD.

Ubuntu Touch, three years later

dplanella — Fri, 18 Dec 2015 15:27:38 +0000

While the article raises some valid points, others have already pointed out that reviewing a retail device or at least use one of the stable channels would give a better impression of the Ubuntu phone experience.

However, being quite honest, I also feel that the tone and some of the statements are a bit unfair and could have benefitted from some more research or reaching out to the Ubuntu community to get the answers: claiming that Canonical hides images because it's not obvious they are available at http://system-image.ubuntu.com, for instance.

In any case, after some of the valid points raised, we updated the image channels page to make clear the difference between stable and devel channels and explicitly not recommend 'devel' unless needed for a specific purpose.

https://developer.ubuntu.com/en/start/ubuntu-for-devices/...

Alternatively, some community members have also produced excellent technical articles about the subject:
- https://sturmflut.github.io/ubuntu/touch/2015/05/05/hacki...
- https://sturmflut.github.io/ubuntu/touch/2015/05/06/hacki...

Thanks!

Ubuntu Touch, three years later

oever — Fri, 18 Dec 2015 08:38:26 +0000

I'm in the process of replacing my trusty Nokia N950 with Meego. I picked up two Ubuntu Phone devices as candidates: The 4.5 Aquaris and the Meizu MX4.
Both devices are nice to use. My demands on a mobile phone are not very high. It has to make phone calls, allow me to send messages, possibly browse the web, act as hotspot of provide data to my laptop via USB.
I've written quite some apps for personal use for the N950 and I plan to port them to Ubuntu Phone.
The browser on both devices is very fast.
My main wish for FOSS phones is that they are very easy to modify. The big promise of free software is that the user can inspect the code and change it. This should be possible in practice, not just in theory. So a simple view into the code of the applications with the option to change code and have it magically compiled and stored as new, derivative version of the original app would be a revolutionary feature.

Ubuntu Touch, three years later

jamesh — Fri, 18 Dec 2015 00:54:46 +0000

The development channel really quite a long way from prime time at the moment. The stable and development releases are on opposite sides of the GCC 5 transition, so you lose access to a fair number of apps in the store.

While things are still being landed in devel, images aren't being regularly promoted from devel-proposed to devel:

$ ubuntu-device-flash query --device=mako --list-images --channel=ubuntu-touch/devel/ubuntu
1: description='ubuntu=20141201,device=20141119,custom=20141201,version=1'
2: description='ubuntu=20150413,device=20150210,custom=20150413,version=2'
3: description='ubuntu=20151127,device=20150910,custom=20151127,version=3'

$ ubuntu-device-flash query --device=mako --list-images --channel=ubuntu-touch/devel-proposed/ubuntu
...
375: description='ubuntu=20151215,device=20150910,custom=20151215,version=375'
376: description='ubuntu=20151216,device=20150910,custom=20151216,version=376'
377: description='ubuntu=20151217,device=20150910,custom=20151217,version=377'

So I agree that it the devel channel is not a good place to point people at (although the page you linked seems to instead suggest table).

If you want to see daily updates that will feed into the next OTA release, the rc-proposed channel is probably the place to look. You'll have an image compatible with the apps on the store, with roughly daily updates:

$ ubuntu-device-flash query --device=mako --list-images --channel=ubuntu-touch/rc-proposed/ubuntu
...
315: description='ubuntu=20151215,device=20150911,custom=20151215,version=315'
316: description='ubuntu=20151216,device=20150911,custom=20151216,version=316'
317: description='ubuntu=20151217,device=20150911,custom=20151217,version=317'

If you want to review what actual end users are running, then the stable channel is the place to look (which also looks to currently be more recent than what was tested):

$ ubuntu-device-flash query --device=mako --list-images --channel=ubuntu-touch/stable/ubuntu
...
24: description='ubuntu=20151015,device=20150911,custom=20151015,version=24'
25: description='ubuntu=20151118.2,device=20150911,custom=20151118.2,tag=OTA-8,version=25'
26: description='ubuntu=20151210,device=20150911,custom=20151210,tag=OTA-8.5,version=26'

Ubuntu Touch, three years later

flussence — Thu, 17 Dec 2015 23:53:47 +0000

I've found the absence of swipe-gesture typing painful on a deblobbed Android too — the stock keyboard bizarrely draws gesture trails, but doesn't act on them unless the proprietary Play Services addon is installed. I have yet to find a properly FOSS replacement for the functionality so far.

Ubuntu Touch, three years later

jrjohansen — Thu, 17 Dec 2015 23:50:05 +0000

But that doesn't explain everything. Are you trying to tell me that the slow response, lack of features, poor battery life, ancient kernel, occasionally confusing interface, etc. are all regressions introduced in this development cycle?

So to address the points individually

slow response - partially, there are fixes in the stable channel that are not in the unstable channel. It is quite likely that any bugs/regressions could also have been triggering bug reporting (unstable is a dev channel) which has been known to cause slow downs and stuttering and other such issues. However there are still known performance issues that need to be addressed.

Lack of features - partially, there are features included in commercial devices (ie. ubuntu versions of BQ and Meizu) that can not be supplied in the community images, and depending on when the image was grabbed it is possible that the unstable channel is missing features in the stable channels. But as already mentioned mail, calendaring, and maps are not provided in the stable image at this time.

poor battery life - probably. Regressions and bugs can kill battery life by preventing the device from entering deep sleep and spiking cpu loads. In the testing I have done on the nexus 4 battery life fairly comparable or better than that of android, and the BQ version is generally better.

ancient kernel - this will always be the case for android based devices, due to the need to support device/vendor blobs and drivers. The kernel is based on a recent version of the android kernel released for the device. And then patches are applied on top, to make it a more Ubuntu kernel. Patches for kernel vulnerabilities are applied when available and applicable.

confusing interface - unlikely. I have difficulties with the interface as well, while other people really seem to like it.

Just let them code

ovitters — Thu, 17 Dec 2015 23:08:00 +0000

There's a cycle of increasing the minimum requirement (e.g. memory, storage needs, display requirements, etc) followed by some new small hardware which only supports way less. The focus is nice. Usually the effort isn't strictly separated anyway. Improvements which they need might require them to pay some programmers or upstream some work.

What's funny is the evolution-data-server component (it is pretty badly named). This was one of the first components every written. It has loads of code. It needs lots of cleanup (progressing slowly). In Nokia tablets they added lots of hacks (it wasn't upstreamed so IMO hacks) to speed up evolution-data-server. Now many years later it still needs way more work. So unfortunate. Hopefully they'll sponsor the work well enough so that everything is upstreamable. Better experience on the desktop for the rest of us.

Lastly, I like my Android phone. That said: go free software!

Ubuntu Touch, three years later

bronson — Thu, 17 Dec 2015 20:54:18 +0000

From https://developer.ubuntu.com/en/start/ubuntu-for-devices/... :

The choice appears to be "Run a well tested OS" and "Track the latest development".

Seems rather obvious that the reviewer would want to track the latest development. If life outside stable is as dire as you imply, I hope you'll update that wiki page.

Ubuntu Touch, three years later

bronson — Thu, 17 Dec 2015 20:45:51 +0000

But nobody's worried about the GPS. How about the other stuff mentioned by the esteemed editor?

Ubuntu Touch, three years later

mhall119 — Thu, 17 Dec 2015 16:43:12 +0000

The hardware GPS takes a very long time to get a location lock, and it seems especially bad on the Nexus 4. On Android it will use Google's AGPS (data provided by Ingress players) to get "close enough" location data right away. Nokia HERE provides the same function for Ubuntu Phone, but it's not open sourced so it does not get included in the base images. As somebody mentioned above, it does come included on the commercial phones being sold by Bq and Meizu, and that makes location services much more functional.

Ubuntu Touch, three years later

corbet — Thu, 17 Dec 2015 14:42:57 +0000

No, I don't quite get the analogy.

It is my tendency to test early-stage software; Ubuntu Touch did not get special treatment there. I want to see where a project is going, rather than where it has been. I know to expect some issues in such situations; that's why I didn't dwell, for example, on the failure of GPS to work for me. It was clearly meant to work, and I believe it normally does.

Ubuntu Touch, three years later

simosx — Thu, 17 Dec 2015 13:53:01 +0000

> In particular, your editor tested version 16.04r3 from the testing channel on a Nexus 4 handset.

Wait, what?

The _testing_ channel has different semantics from, let's say Debian Testing.

@corbet: you should have used the "stable" channel.

Ubuntu Touch, three years later

mzanetti1 — Thu, 17 Dec 2015 13:09:19 +0000

It's a bit sad that those articles are most of the times written after not doing enough research. Taking the testing channel on a Nexus 4 and complaining about issues is a bit like reading the draft of the free/teaser version of a LWN article and telling people that LWN articles lack information and have typos. Which is obviously not true, but I hope you get the analogy.

Ubuntu Touch, three years later

popey — Thu, 17 Dec 2015 12:19:30 +0000

It's interesting to me to see feedback from people outside the project, especially from those who haven't used Ubuntu on a mobile device for a while.

It's a shame you'd used the unstable channel rather than the well tested and stable channels also available. I think your experience might have been a touch better, especially with regards to sluggish behaviour. I do wonder what directed you to do that, and perhaps there's some guide / documentation we could improve there.

The weather app isn't "limited to a few dozen hardwired locations worldwide". It comes with some pre-defined, but you can search (magnifying glass icon) for any additional location supported by the data providers (The Weather Channel & OpenWeatherMap).

http://imgur.com/a/cpfMi

Terminal asks for a passcode/phrase because it's an unconfined application and as such could cause some damage if used inappropriately. It's there so someone (a mean 'friend') playing with the your phone can't do anything catastrophic (like rm ~/*) without knowing your pin.

For the IMAP client, we have one in development which will land in default images once it's ready.

Clarifications

rlqb — Thu, 17 Dec 2015 11:11:35 +0000

OSM Scout is an offline map app that I think renders from vector data (I have the whole of the UK stored on it). It is still very primitive (almost no features) and so is nothing like OsmAnd but it's a start. It works well enough for just viewing a map of the current location, anyway.

Clarifications

tajyrink — Thu, 17 Dec 2015 11:00:49 +0000

Some potentially useful bits of information for the editor and others follows. I've noticed the traditional desktop/server distro mindset sometimes leads to misunderstandings or wrong choices.

It's recommended to use the "stable" (or "rc-proposed" for daily images) channels instead of devel channels for everyone who is not interested in fixing broken pieces. There are key differences to desktop in the Ubuntu phone images compared to traditional distros. Practically all user visible new features do appear in the stable channel too. The devel is more about coping with the base distribution toolchain and library upgrades etc, meaning that devel channel more likely than not is having many things broken because all current customers are running on the stable channel that's based on Ubuntu 15.04 (+ all the development after it and security updates) at the moment. Focus of all development is there while 16.04 is tried to keep in sync and in the future stabilized again.

The images are available at http://system-image.ubuntu.com/ - a bit similar to Debian archive organization, there are indexes and then a pool of images. Each device has a device, ubuntu and custom images which are used together.

To compare to Android product I'd compare an Ubuntu product - even though Nexus 4 port is supported by Canonical it's not a product in the same sense as what one finds in stores. Granted the US availability still limits the buying option of eg Bq for people around there. But I find that Bq performs pretty well for its price range, has an excellent browser and getting better with every OTA upgrade.

Some shortcomings are certainly very true. Dekko is a capable e-mail client but (still) not included in the default installation. I recommend installing it anyway, or if one just uses GMail there's an own app for it. Calendar is in similar situation, it has not been selected in the default installation until certain bugs are fixed first. For mapping application, the sold products come with HERE maps + assisted GPS functionality which obviously can't be included in the community version because of licensing. There's not yet a swiss knife of free mapping application that I'd compare to eg OsmAnd. uNav is pretty good however in that it offers voice guided navigation and is constantly updated - but I'd want a vector graphics based application like Navit because otherwise offline maps are unfeasible.

The app startup time is not yet as good as it could be, but luckily there are proof of concepts to fix that.

Ubuntu Touch, three years later

habarnam — Thu, 17 Dec 2015 10:57:28 +0000

Sorry for moving a little off topic, but I would be interested in Mr. Corbet's opinions on testing Sailfish OS on the same device.

As far as I know it's a lot more mature, despite the troubles Jolla has been facing in the recent future. And based on the article, the end product is a lot more linuxy than Ubuntu Touch.

Ubuntu Touch, three years later

pauldub — Thu, 17 Dec 2015 08:57:57 +0000

I have to second your comment, I own a bq Aquaris E4.5 and have none of the memory issues and the battery is just fine (the phone is up for several days, usually 2-3). Admittedly I do not use many applications and I think that sometimes the phone will not go into powersaving mode because of an application but overall it's really smooth.

Ubuntu Touch, three years later

JanC_ — Thu, 17 Dec 2015 08:19:54 +0000

I'm using a bq Aquaris E4.5 that came with Ubuntu here (running 15.04 OTA-8.5 now) and it seems like I don't have *all* the issues you are seeing, but there certainly still are bugs there (e.g. in the calendar app…).

I usually don't get into memory problems, unless I open a lot of webpages (including "webapps"). We can blame Chromium's stupid one-huge-process-per-page model for that, I guess (which manages my 16 GiB RAM desktop to start swapping too!), as that's what Oxyde is based on.

A short swipe from the right edge switches to the previous application, a long swipe (over halfway) opens the application overview. In the application overview you can select another application or close applications (by swiping their window up or down off the screen).

The reason why the terminal app requires a password (or PIN code, depending on your configuration) is that it is not strictly confined with AppArmor like other apps are. Normal apps can't access anything on the filesystem outside their own "private space", and that would be unpractical for the terminal. (The file browser also needs authentication if you want to access files outside a certain area.)

For IMAP you should be able to use the (3rd party) Decko app in the store.

BTW: it seems like you are using a development image, which can possibly "explain" some of the issues you see.

Just let it die

b7j0c — Thu, 17 Dec 2015 04:27:02 +0000

Totally hopeless.

Ubuntu Touch, three years later

khim — Thu, 17 Dec 2015 01:29:08 +0000

I think Ubuntu Touch's claim to fame was going to be that you can run the same OS on a phone or on a PC (is that true?).

Not “the same OS on a phone or on a PC”… “the same device could be used as phone or PC”. Ubuntu Edge-likes. And they will come. But without Ubuntu.

Hardware is just not there: you need SOC with CPU and GPU powerful enough to drive desktop apps and contemporary phone SOCs while powerful are just not powerful enough. But we'll arrive there in 3-5 years.

That's why Apple isn't really concerned about it.

That's because Steve Jobs is no longer with us. He would have tried to really strain Apple's resources to bring the “the unified world” 3-5 years before competitors. Yes, that's tough, yes, this means that you'll stop making as much money… which is why accountants like Tim Cook couldn't do that. But that just means that Apple, instead of leading the revolution, will slowly start slipping into the irrelevance. The rule “if Apple does not do X then X is not the future” is no longer working.

one has to really make an effort to distinguish why your particular mobile OS is better, and I mean a *lot* better than what's already out there (and already popular).

Too late. The war is already lost. Smartphone penetration rate have passed 50% in many countries and even laggards like India are moving in that direction. That means that future smartphone buyers from now on are mostly current smartphones owners: they know the UI (and would look for similar one), they own tons of programs (except for the ones who use smarthone as a dumbphone… but it's highly unlikely that such users will want to use their phones as a PC), etc. Ubuntu, Tizen and others have run out of time—it's as simple as that. Android no longer need any revolutions too keep that market for itself. Unless Google will do some kind of major blunder (similar to craziness Microsoft did with switch from Windows Mobile to Windows Phone 7 to Windows Phone 8 to Windows 10 Phone) it's future as the next king of desktop is pretty much guaranteed… even if Google itself does not believe it (and it does not: I have many friends from Google and they call me crazy when I say that in 5-7 years Android would own the desktop, not Windows)…

Ubuntu Touch, three years later

k8to — Thu, 17 Dec 2015 00:54:42 +0000

I just expected this to end in failure due to misplaced confidence. Companies who realize their thing can be repurposed or adapted to do another thing really well sometimes have massive success. Companies who decide their big complex thing already is "pretty close" to some other thing seem to always fail.

I just couldn't imagine a linux distribution becoming a top quality tablet OS in a short timescale. The whole thing smelled of hubris.

Ubuntu Touch, three years later

marduk — Wed, 16 Dec 2015 21:09:35 +0000

It's starting to smell like Microsoft's Kin. I also think that, given iOS and Android (and even Blackberry OS, Windows Mobile & Tizen) one has to really make an effort to distinguish why your particular mobile OS is better, and I mean a *lot* better than what's already out there (and already popular).

I haven't seen anything indicating that from Ubuntu Touch or Firefox OS. What are you giving me that I don't already have? Bonus points if it's something I actually care about.

I think Ubuntu Touch's claim to fame was going to be that you can run the same OS on a phone or on a PC (is that true?). But it seems that the market doesn't really care about that much. That's why Apple isn't really concerned about it. That's why Windows Mobile hasn't really gone anywhere. So if people aren't concerned about it then you have to have something else that's going to attract people that the others don't have. The "free" thing is only going to work for a very small subset of the market and, to be frank, for me to equate "Canonical" with "free" is not an easy pill to swallow. I'd probably go more to Firefox OS if that was what I was going for.

Recognition to Mr. Corbet for an article which was as entertaining as it was informative.